Originally published here at TechCrunch

Companies are experimenting with services like ChatGPT to help workers write memos, answer questions, and much more. Yet, one of the clearest signs of how much traction AI is really getting in the world of enterprise IT is how often it figures in the more routine applications that organizations use. Today, a startup that exemplifies that axiom is announcing funding, underscoring that pace of enterprise AI adoption.

Opkey has built an AI platform to help organizations continuously test finance, HR and other enterprise resource planning (ERP) software. Now, on the back of some strong customer traction — more than 200 large enterprise customers and partnerships with system integrators like KPMG and PwC — the startup has closed a $47 million Series B.

PeakSpan Capital is leading the round, which saw participation from previous backers and its current investors, including UST Global, Verica, Vertical and India’s YouNest. This Series B is a big step up for Dublin, California-based Opkey: The company had only raised $12 million before now. The startup is not disclosing its valuation.

Opkey’s platform may not be at the most visible end of the enterprise IT stack, but it’s addressing a gap in the market that is critical to how enterprises operate.

Cloud architecture and SaaS have become the bedrock on which new business services are built, but they have also found a lot of footing with more legacy organizations undergoing digital transformation processes. They can be faster and cheaper to use and easier to deploy to a user base, but they can also come with a ton of problems — like the fact that any vulnerabilities or inconsistencies between software can take down entire networks. This is where Opkey comes into the picture.

Pankaj Goel, the CEO, co-founded the company with his childhood friends Avinash Tiwari and Lalit Jain. All three are ERP industry veterans with experience at big companies like Adobe and Oracle. Their time there gave them experience in software testing as well as a front-row seat to how disastrous it could be if not done correctly. ERP systems typically do not exist in silos — they integrate with each other to work, which means if one conflicts with another, or is not working correctly, the whole ERP stack can crash.

“Cloud applications are continuously giving updates, which was not the case with traditional software,” Goel said in an interview. “That functionality breaks existing functionality.” He estimates that an organization might typically integrate seven or eight ERP systems together. “Any change in the ecosystem, and you need to test it again. It means enterprises are in testing hell.”

Testing and data hell are in fact classic problems AI-based automation can resolve, and cybersecurity and DevOps software companies have also come to realize that. In the case of Opkey and ERP, the platform continuously tracks integrations, updates, upgrades, and something the company describes as “user acceptance” — how well end users are interacting with new features they are supposed to be using. Packages and platforms that Opkey currently covers include Oracle, Workday, Coupa, Veeva, Salesforce, SAP, Microsoft Dynamics, UKG and Trackwise.

Sanket Merchant, the partner at PeakSpan who led the investment, believes that there will continue to be a very strong funnel of business interest for services like Opkey’s given the direction IT is moving.

One question is if that will be enough to stave off competition from others in the same space, such as Leapwork and Katalon. Whether bigger ERP players may look to take on this business themselves is another.

“For businesses, the most expensive IT spend is related to critical enterprise apps, their ERP apps,” he said, citing figures that say that some $73 billion is spent every year on ERP software for billing, accounting, people management, software deployment and more across both smaller businesses and massive, multinational organizations. These in turn can have as many as 52 different other apps that rely on ERP integrations to work.

“Automated testing is important to drive assurance around that investment, to make sure it behaves as intended. A failed SAP deployment or Workday migration can have a huge impact on a business’s revenue and its brand.”